Attention deficit hyperactivity disorder recognition system based on brain function connection group development norm

Abstract

The invention provides an attention deficit hyperactivity disorder recognition system based on a brain function connection group development norm, which is used for classifying and recognizing attention deficit hyperactivity disorder by constructing the brain function connection group development norm and utilizing an obtained brain function connection individual development deviation index. The invention provides an attention deficit hyperactivity disorder recognition system based on a brain function connection group development norm. According to the method, a norm of brain function connection characteristics changing along with age is constructed, the degree of an individual deviating from a normal range is described, and the attention deficit hyperactivity disorder patient is identified based on an obtained individualized deviation index, so that a quantitative and individual-level novel brain imaging biological marker is provided for the attention deficit hyperactivity disorder patient. In norm construction, the Bayesian spline regression method used in the method can more accurately model nonlinear distribution by combining B spline transformation, and meanwhile, the calculation time can be shortened, so that the method can be efficiently expanded to a large brain image data queue and can be flexibly applied to new sites which are not included in training samples.

Description

Technical Field

[0001] The present invention relates to an attention deficit hyperactivity disorder recognition system based on the developmental norm of brain functional connection groups. Background Art

[0002] Attention-deficit / hyperactivity disorder (ADHD) is one of the most common neurodevelopmental disorders in children and adolescents. Its core symptoms are inattention, impulsivity, and hyperactivity that are inconsistent with age. Currently, the diagnosis of ADHD mainly relies on clinical evaluation and behavioral observation, which is highly subjective. Therefore, it is urgent to reveal objective and reliable biomarkers to achieve early identification and accurate diagnosis and treatment of ADHD.

[0003] CN202310853785.6, Invention Name: A method and system for assisting diagnosis of attention deficit hyperactivity disorder, providing a method and system for assisting diagnosis of attention deficit hyperactivity disorder, and relating to the field of medical image processing and assisting diagnosis. The method for assisting diagnosis of attention deficit hyperactivity disorder comprises: dividing the acquired brain time series data into at least one continuous and non-overlapping sub-time window; extracting the discriminative features of each sub-time window, and obtaining a high-order functional connection network through the extracted discriminative features; capturing the dynamic features of the acquired high-order functional connection network of each sub-time window through the TCN module, and characterizing the relationship between brain functional connections over time; obtaining the value of the correlation between each sub-time window and the disease through the dynamic change law of brain diseases, and using the value of the correlation between each sub-time window and the disease as a weight to perform feature fusion on the high-order functional connection network of each sub-time window. The invention uses dynamic functional connections for assisting diagnosis of ADHD. In order to capture the evolution characteristics of functional connections over time, dynamic functional connections usually require longer scanning times and higher time resolutions to obtain more accurate and representative dynamic connection patterns, and have a high dependence on scanning time, which increases the difficulty of data acquisition. In addition, dynamic functional connectivity is more sensitive to noise. Given that dynamic functional connectivity is easily affected by scanning time, temporal resolution, and noise, the equipment and operating environment of different scanning centers lead to different noise levels and different scanning time parameter settings, which may lead to a decrease in the generalization ability of the model.

[0004] CN202311304937.3, Invention Name: An auxiliary diagnosis system for attention deficit hyperactivity disorder syndrome based on a comprehensive cognitive experiment set and a multi-stream dual-row decision neural network model, provides an attention deficit hyperactivity disorder syndrome reaction test system. The reaction test system is composed of multiple cognitive reaction experiments, all of which are experiments designed and optimized based on cognitive psychology and cognitive neurology data. A comprehensive assessment of the subject is achieved. During the experiment, the EEG data of the subject is collected, and then the frequency band of interest in the reaction experiment is processed according to the EEG data of the experiment. The data conversion and feature extraction are sent to the multi-stream dual-row decision neural network (MSDD) model, which selects different learning schemes according to different data types, and learns the EEG data features to compare their similarity with the healthy norm, so as to achieve the purpose of auxiliary diagnosis. The invention uses EEG data to identify ADHD. The spatial resolution of EEG is much lower than that of functional magnetic resonance imaging (fMRI), and it is difficult to accurately locate the activity source inside the brain. In addition, EEG signals are easily affected by myoelectric interference (such as blinking, facial muscle activity, etc.) and external electromagnetic interference, and complex signal processing technology may be required to purify the data.

[0005] CN201910987452.6, Invention Name: An Attention Deficit Hyperactivity Disorder Auxiliary Diagnosis Device and Its Use Method, proposes an Attention Deficit Hyperactivity Disorder Auxiliary Diagnosis Device and Its Use Method, including: a display screen for interacting with the subject and displaying the test scene; a wearable accelerometer for collecting the limb acceleration of the subject in the test scene; a data processing module for preprocessing the limb acceleration to obtain the test data, inputting the test data into the Attention Deficit Hyperactivity Disorder detection model based on a deep neural network for identification, obtaining the auxiliary diagnosis result, and sending it to the display screen. The invention uses a wearable accelerometer to measure limb acceleration to identify ADHD, but the motion data of the accelerometer may be affected by many factors, such as the subject's state (such as anxiety, fatigue, excitement, etc.), motion errors, and some other neuropsychiatric diseases will also show abnormal movements, such as bipolar disorder, schizophrenia, tics, etc., so its specificity for auxiliary diagnosis of ADHD may be insufficient.

[0006] CN202010922095.8, Invention Name: ADHD Disease Diagnosis Decision-Assisting System Based on Deep Convolutional Spiking Neural Network, discloses an attention deficit hyperactivity disorder (ADHD) disease diagnosis decision-assisting system based on deep convolutional spike neural network, including: a preprocessing device for preprocessing functional magnetic resonance imaging data using a signal-to-noise ratio feature selection method; an encoding device for encoding the preprocessed functional magnetic resonance imaging data using a forward algorithm to generate a pulse sequence; a diagnosis prediction output device for using a deep convolutional spike neural network to perform a diagnosis prediction based on the pulse sequence and output a diagnosis prediction result. This invention uses a pulse sequence generated by raw functional magnetic resonance imaging data to perform diagnostic decision-assisting for ADHD. The data dimension is very high, and the processing and analysis are extremely complex. In addition, the interpretability of the indicators is poor, and the physiological significance is unclear.

[0007] In addition to the above-mentioned differences in observation indicators between the present invention and several other inventions, all previous inventions build machine learning classification models based on observation data, and make ADHD judgments based on the diagnostic results output by the classification model. However, machine learning models, especially deep learning models, are often regarded as "black boxes" and it is difficult to explain their decision-making process. In the medical field, especially in mental health diagnosis, clinical practitioners need to understand the decision-making basis of the model, and this unexplainability hinders its application. Functional connectome analysis technology based on resting-state functional magnetic resonance imaging is considered to be the most effective means to reveal the brain functional mechanisms of various mental illnesses and has been widely used in ADHD, such as Sripada, C., et al. Disrupted network architecture of the resting brain in attention-deficit / hyperactivity disorder. Human brain mapping 35, 4693-4705 (2014); Norman, LJ, Sudre, G., Price, J., Shastri, GG & Shaw, P. Evidence from "big data" for the default-mode hypothesis of ADHD: a mega-analysis of multiple large samples. Neuropsychopharmacology 48, 281-289 (2023); Norman, LJ, Sudre, G., Price, J. & Shaw, P. Subcortico-Cortical Dysconnectivity in ADHD: AVoxel-Wise Mega-Analysis Across Multiple Cohorts. The American journal of psychiatry 181,553-562(2024).

[0008] The above studies have explored the differences in brain functional connectivity group characteristics between ADHD patients and normal development control groups, and found that ADHD patients have abnormal functional connectivity in the default mode network, frontoparietal network, salience network, etc. However, the results between studies are highly inconsistent and cannot be translated into clinical applications. Because ADHD is a highly heterogeneous disease, different patients have great differences in clinical symptoms and potential biological characteristics of the disease. The above studies all used the traditional case-control comparison method, treating patients with the same diagnosis as a homogeneous group. Therefore, the results obtained reflect the differences in group averages, while ignoring the heterogeneity between individual patients, and also ignoring the impact of development on brain function. This group-level difference is not generalizable, and different results are obtained by comparing and analyzing a different group of patients, so it cannot be used for the accurate identification of individual ADHD patients.

[0009] The emergence of normative modeling methods provides a new approach to solving this problem. This method constructs a norm of brain structure or functional characteristics based on brain imaging data of a large sample of healthy people. The obtained norm can reflect the typical pattern and variation reference range of brain structure or function under specific age, gender and cognitive status, so it can be used to assess the degree of deviation of individuals from the normal pattern and "microscopically" reveal the differences of each patient at the individual level, which is of great significance for the early diagnosis and identification of patients with mental disorders and precise treatment.

[0010] Currently, there are no norms for the development of brain functional connectomes for children and adolescents, and the constructed norms for the development of brain functional connectomes have not been used to identify patients with attention deficit hyperactivity disorder. Summary of the invention

[0011] The present invention proposes an attention deficit hyperactivity disorder identification system based on the developmental norms of brain functional connectomes. By constructing a reference model of how the characteristics of brain functional connectomes change with age in a normally developing population, the degree to which each ADHD patient deviates from the normal range can be obtained based on this model, and an individualized brain functional connection deviation pattern can be obtained. Based on the deviation pattern, ADHD patients and normally developing controls are classified and distinguished, thereby achieving identification of ADHD patients.

[0012] The present invention provides an attention deficit hyperactivity disorder identification system based on the brain functional connection group development norm, which constructs the brain functional connection group development norm and uses the obtained brain functional connection individual development deviation index to classify and identify attention deficit hyperactivity disorder.

[0013] Wherein, the brain functional connection group development norm is constructed by using Bayesian spline regression.

[0014] The attention deficit hyperactivity disorder recognition system based on the brain functional connection group development norm of the present invention includes the following functional modules:

[0015] a. Data acquisition module: used to collect demographic and magnetic resonance imaging data of subjects;

[0016] b. Data preprocessing module: connected to the data acquisition module, used for correction, registration, noise reduction, smoothing and filtering preprocessing of the collected resting state functional magnetic resonance imaging data;

[0017] c. Brain functional connection group calculation module: connected to the data preprocessing module, generating a functional connection matrix from the preprocessed resting state functional magnetic resonance imaging data, and calculating the functional connections of brain regions and brain networks;

[0018] d. A brain functional connectome developmental norm construction module: connected to the brain functional connectome calculation module, in the normal development control group, using Bayesian spline regression to construct the norm of each brain region and brain network functional connection characteristics with age, based on the developmental norm, calculate the degree of deviation of each functional connection characteristic of each attention deficit hyperactivity disorder patient from the norm, and obtain the individual developmental deviation score;

[0019] e. Attention Deficit Hyperactivity Disorder (ADHD) Identification Module: It is connected to the brain function connection group development norm construction module, and according to the obtained brain function connection development deviation score, a support vector machine algorithm is used to construct a classification model for ADHD patients and normal development controls, and the brain function connection pattern with individual extreme positive deviation and negative deviation is combined to determine whether the patient is an ADHD patient.

[0020] The data acquisition module is used to collect demographic information, T1 structural magnetic resonance imaging and resting-state functional magnetic resonance imaging data of multi-center ADHD patients and normal development control subjects;

[0021] The data preprocessing module is for image data preprocessing. The functional magnetic resonance imaging data of all data sets are preprocessed using the brain imaging data processing and analysis toolbox based on the MATLAB platform;

[0022] Brain functional connectome calculation: used to analyze and understand the functional coordinated activities or connection patterns between different brain regions, calculate the correlation of time series between different brain regions and networks, and obtain brain functional connectome characteristics;

[0023] The developmental norms of brain functional connectomes were constructed in a control group with normal development. The data set was divided into 10 parts using a stratified 10-fold cross-validation method. The proportion of data from each center in each part was the same as that of the entire data set. One part was selected as the validation set each time, and the remaining 9 parts were used as the training set, and this cycle was repeated 10 times. Males and females were modeled separately to reduce the impact of gender differences on the model. Based on the training set data, Bayesian Spline Regression was used to characterize the changing pattern of functional connectivity characteristics of each brain region and brain network with age. Bayesian Spline Regression introduced cubic B-spline transformation into the general Bayesian linear regression model to capture nonlinear effects, and the Powell method was used to fit and optimize hyperparameters. Explained variance (EV) and mean standardized log loss (MSLL) were used to evaluate the norms.

[0024] The individual deviation score calculation method is as follows: based on the established brain functional connectivity norm, the degree to which each brain functional connectivity feature d of each subject n deviates from the norm is statistically estimated, that is, the individual deviation score Z nd , and its calculation formula is:

[0025]

[0026] Among them, y^ nd is the predicted functional connectivity value, y nd is the true functional connection value, using noise σ 2 d Variance (reflects variability in the data) and uncertainty variance (σ 2 * ) d Normalize.

[0027] Furthermore, the image data preprocessing method includes: removing the first 5 time points; time layer correction; head motion correction; image registration to T1 structural image and then to MNI standard space; regression Friston-24 head motion parameters, linear drift, white matter signal, cerebrospinal fluid signal and whole brain average signal; smoothing with a 6mm Gaussian smoothing kernel; 0.01-0.1Hz bandpass filtering.

[0028] The method for calculating the brain functional connectome comprises the following steps:

[0029] S1: The cerebral cortex was divided into 400 regions of interest (ROIs) using the Schaefer brain function atlas;

[0030] S2: Calculate the Pearson correlation coefficient between the time series of these 400 ROIs and perform Fisher-z transformation to obtain a 400×400 functional connectivity matrix;

[0031] S3: Calculate the average functional connectivity value of each ROI [the sum of the functional connectivity between the ROI and other ROIs / (total number of ROIs-1)] to obtain 400 brain region-level functional connectivity values;

[0032] S4: According to the Yeo brain functional network atlas, these 400 ROIs were assigned to 7 brain networks, and the functional connectivity within and between networks was calculated, with a total of 28 [7+7×(7-1) / 2] functional connectivity values; the intra-network functional connectivity of each brain network was the average of the functional connectivity between the ROIs within the network, and the inter-network functional connectivity was the average of the functional connectivity between the ROIs of two brain networks.

[0033] The steps of constructing the norm are as follows:

[0034] (1) The normal control data set is divided into 10 parts, and the proportion of data from each center in each part is the same as that of the entire data set. Each time, one of the parts is selected as the validation set, and the remaining 9 parts are used as the training set. This cycle is repeated 10 times. Based on the training set data, a cubic B-spline basis function expansion is used for the independent variable age. φ(x) represents this expansion, where Φ is the N x K matrix containing the basis function expansion of all subjects; the single functional connection value y is assumed to be a linear combination of the B-spline basis function transformation plus the noise term:

[0035] y=w T φ(x)+∈ s

[0036] where w is the estimated weight vector, ε s is the Gaussian noise distribution of site s, with mean 0 and precision β s ;

[0037] All noise accuracies from different sites will be combined in a vector β, allowing the noise accuracies to vary between sites to accommodate inter-site variability, as well as site-specific intercepts; dummy variable site regressors are added to the final basis expansion matrix of the B-spline to make it a random intercept model;

[0038] (2) Use Bayesian linear regression to give model parameters Set a Gaussian prior, where α is a weight-dependent hyperparameter; assume that the mean of the Gaussian prior is zero and the precision matrix is ​​Λ α , using the isotropic precision matrix Λ α = αI; the Gaussian prior choice allows the calculation of the posterior distribution of w in closed form:

[0039]

[0040] The posterior for each subject was obtained using standard posterior derivation:

[0041]

[0042] The Powell method is used to fit the optimal hyperparameters α and β; Powell is a derivative-free method that can fit hyperparameters faster. The resulting predicted distribution is:

[0043]

[0044] This predicted distribution reflects the change pattern of brain functional connectivity characteristics with age, thereby obtaining the developmental norms of each brain region and network functional connectivity. The model was evaluated using the above validation set data, and evaluation indicators such as explained variance (EV) and mean standardized log loss (MSLL) were calculated to evaluate the fitting effect of the norm.

[0045] The method for identifying attention deficit hyperactivity disorder described in step e specifically comprises the following steps:

[0046] S1: Using patients with attention deficit hyperactivity disorder as a test set, predicting the brain functional connectivity values ​​of patients with attention deficit hyperactivity disorder based on the above-constructed brain functional connectivity norm, calculating the individual deviation score of each brain functional connectivity feature of each ADHD patient, defining the absolute value of z greater than 2 as an extreme deviation, in order to identify brain regions or networks showing extreme positive and negative deviations;

[0047] S2: Combine the deviations of all brain regions and network functional connectivity to analyze the overall deviation pattern of ADHD patients, and use extreme value statistics to analyze the most extreme deviation brain regions and networks of individuals;

[0048] S3: The individual deviation scores of brain functional connectivity between patients with ADHD and normal development controls were used as input features to construct a support vector classification model to discriminate patients with ADHD.

[0049] The feature adopted by the present invention is a static functional connection group calculated based on resting state functional magnetic resonance imaging (fMRI) data. The resting state functional magnetic resonance imaging measurement environment adopted is more controllable, the subject only needs to remain motionless, the degree of cooperation is higher, the data acquisition is relatively more stable, and the obtained brain functional connection group characteristics can provide a potential more specific neural activity pattern, which helps to improve the accuracy of diagnosis. The static functional connection analysis adopted is the average connection pattern during the entire scanning process, which requires relatively less scanning time, is relatively less sensitive to noise, and is more stable, so the generalization in different scanning centers is better. The present invention calculates the brain functional connection group index based on fMRI data, which reflects the synchronization of functional activities between different brain regions or brain networks, integrates complex information into a relatively easy-to-understand connection pattern, reduces the redundancy of data, improves the efficiency of information extraction, is more feasible as a biomarker, and is more practical in clinical applications.

[0050] The present invention constructs a norm and calculates the degree to which each individual's brain function characteristics deviate from the norm, thereby obtaining the brain area or network of each individual's abnormal function, and issuing a brain function imaging report similar to the traditional diagnostic report of the radiology department to clinicians, providing clinicians with a more interpretable diagnostic tool, thereby improving the accuracy of diagnosis. By establishing a normal model of brain function connection group development in a healthy group, the degree to which each individual deviates from the norm can be calculated, and abnormal individuals can be identified based on whether the individual's degree of deviation exceeds the normal range, and further based on the individual's deviation pattern, auxiliary diagnosis of attention deficit hyperactivity disorder can be performed, thereby achieving a technical upgrade from group average comparison to individual accurate evaluation, and solving the bottleneck problem of clinical application of brain imaging indicators.

[0051] The beneficial effects of the present invention are:

[0052] The present invention provides an attention deficit hyperactivity disorder identification system based on the developmental norms of brain functional connectomes. By constructing a norm in which brain functional connectivity features change with age, the degree to which an individual deviates from the normal range is characterized, and the identification of patients with attention deficit hyperactivity disorder is achieved based on the obtained individualized deviation index, thereby providing a quantitative, individual-level new brain imaging biological marker for patients with attention deficit hyperactivity disorder. In the norm construction, the Bayesian spline regression method used in the present invention can more accurately model nonlinear distributions by combining B-spline transformation, and can also shorten the calculation time, so it can be efficiently expanded to large brain imaging data queues and flexibly applied to new sites not included in the training samples. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a schematic diagram of the process of the present invention;

[0054] Figure 2A schematic diagram of a developmental norm constructed based on the characteristics of the brain functional connectome in an embodiment of the present invention;

[0055] Figure 3 This is a diagram of brain regions with extreme deviations between ADHD patients and normal development controls in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] Example 1 The schematic diagram of the process of constructing the attention deficit hyperactivity disorder recognition system based on the brain functional connection group development norm of the present invention is as follows Figure 1 shown.

[0057] 1. Data Collection

[0058] Demographic data, T1 structural MRI, and resting-state functional MRI data were collected from the ADHD-200 public dataset and the SCU dataset.

[0059] 2. Image data preprocessing

[0060] The functional magnetic resonance imaging data of all data sets were preprocessed using a toolbox for Data Processing & Analysis for Brain Imaging (DPABI, version v4.2) based on the MATLAB platform, using a unified preprocessing process, including: removing the first five time points; time layer correction; head motion correction; functional image registration to T1 structural image and then to the standard MNI space; regression of Friston-24 head motion parameters, linear drift, white matter signal, cerebrospinal fluid signal and whole brain average signal; smoothing with a 6mm Gaussian smoothing kernel; and 0.01-0.1Hz bandpass filtering.

[0061] 3. Brain functional connectome calculation

[0062] The Schaefer brain function atlas was used to divide the cerebral cortex into 400 regions of interest (ROIs). The Pearson correlation coefficients between the time series of the 400 ROIs were calculated, and Fisher-z transformation was performed to obtain a 400×400 functional connectivity matrix. The average functional connectivity of each ROI was calculated [the sum of the functional connectivity of the ROI with other ROIs / (total number of ROIs-1)] to obtain 400 functional connectivity values ​​at the brain region level. According to the Yeo brain functional network atlas, the 400 ROIs were assigned to 7 brain networks, and the intra-network and inter-network functional connectivity were calculated, with a total of 28 [7+7×(7-1) / 2] functional connectivity values; the intra-network functional connectivity of each brain network was the average functional connectivity between the ROIs in the network, and the inter-network functional connectivity was the average functional connectivity between the ROIs of two brain networks.

[0063] 4. Developmental norm construction

[0064] In the control group with normal development, Bayesian spline regression was used to characterize the change patterns of functional connectivity characteristics of the above 400 brain regions and 28 brain networks with age, so as to construct the development norm of brain functional connectivity group. Males and females were modeled separately to reduce the impact of gender differences on the model. The model was verified by stratified 10-fold cross validation. The specific steps are as follows:

[0065] (1) The normal control data set is divided into 10 parts, and the proportion of data from each center in each part is the same as that of the entire data set. One of the parts is selected as the validation set each time, and the remaining 9 parts are used as the training set. This cycle is repeated 10 times. Based on the training set data, a cubic B-spline basis function expansion is used for the independent variable age. φ(x) represents this expansion, where Φ is the N x K matrix containing the basis function expansion of all subjects. The single functional connection value y is assumed to be a linear combination of the B-spline basis function transformation plus the noise term:

[0066] y=w T φ(x)+∈ s

[0067] where w is the estimated weight vector, ε s is the Gaussian noise distribution of site s, with mean 0 and precision β s All noise accuracies from different sites will be combined in a vector β, and we allow the noise accuracies to vary across sites to accommodate inter-site variation, as well as site-specific intercepts (i.e., dummy variable site regressors in the design matrix). The dummy variable site regressors are added to the final basis expansion matrix of the B-spline, making it a random intercept model. This approach provides an efficient way to accommodate site effects in regularized modeling.

[0068] (2) Use Bayesian linear regression to give model parameters Set a Gaussian prior, where α is a weight-dependent hyperparameter. Assume that the mean of the Gaussian prior is zero and the precision matrix is ​​Λ α , where we use a simple isotropic precision matrix Λ α = αI. The choice of Gaussian prior allows the calculation of the posterior distribution of w in closed form:

[0069]

[0070] The posterior for each subject was obtained using standard posterior derivation:

[0071]

[0072] Use the Powell method to fit the optimal hyperparameters α and β. Powell is a derivative-free method that can fit hyperparameters faster. The resulting predicted distribution is:

[0073]

[0074] This predicted distribution reflects the change pattern of brain functional connectivity characteristics with age, thereby obtaining the developmental norms of each brain region and network functional connectivity. The quantile model form is selected, and each model fits seven curves of 5th quantile, 10th quantile, 25th quantile, 50th quantile, 75th quantile, 90th quantile and 95th quantile to reflect the range of changes in brain functional connectivity. Figure 2 Schematic diagram of the developmental norms of functional connectivity of two of the brain regions.

[0075] (3) The model was evaluated using the validation set data. Evaluation indicators such as explained variance (EV) and mean standardized log loss (MSLL) were calculated to assess the fitting effect of the norm.

[0076] (4) In order to quantify the degree of deviation of each subject from the normal range, the deviation z score of each brain functional connectivity feature d of each subject n was calculated as follows:

[0077]

[0078] Among them, y^ nd is the predicted functional connectivity value, y nd is the true functional connectivity value, using the noise variance σ 2 d (reflecting the variability in the data) and the uncertainty variance (σ 2 * ) d The z-score quantifies how much an individual's functional connectivity feature estimate deviates from the model prediction given the model uncertainty.

[0079] 5. Identification of ADHD patients

[0080] Using the ADHD data as the test set, the brain functional connectivity values ​​of ADHD patients were predicted based on the developmental norms of the brain functional connectivity group constructed above. The individual deviation score z of each brain functional connectivity feature of each ADHD patient was calculated, thereby obtaining the degree to which the functional connectivity of each brain region or network of the individual ADHD patient deviated from the normal range. The absolute value of z greater than 2 was defined as an extreme deviation to identify brain regions or networks that showed extreme positive and negative deviations. The deviations of the functional connections of all brain regions and networks were combined to analyze the overall deviation pattern of ADHD patients, and the extreme value statistics method was used to analyze the individual brain regions and networks with the most extreme deviations. The extreme deviation brain regions of ADHD patients and normal development controls are shown in the figure below. Figure 3 As shown in Figure 3, there are almost no extreme deviations in the brain regions of normal development controls, while ADHD patients have widespread positive deviations, that is, the functional connectivity of brain regions is higher than the normal range.

[0081] Finally, the individual deviation scores of brain functional connectivity of all subjects were used as input features, and the support vector machine algorithm was used to construct a classification model for ADHD patients and normal development controls. The nested 10-fold cross-validation method was used to adjust the parameters and test the model. The area under the receiver operating characteristic (ROI) curve (Area Under Curve, AUC), accuracy, sensitivity and specificity were used to evaluate the classification performance of the model. The results showed that compared with the original functional connectivity value, the use of individual deviation scores of brain functional connectivity as features significantly improved the classification performance. The average classification AUC, accuracy, sensitivity and specificity of the classification model in 10-fold cross-validation and their standard deviation values ​​are shown in Table 1.

[0082] Table 1. Performance metrics of ADHD diagnostic classification models

[0083]

[0084] Based on the above-constructed brain functional connection group norm, the degree to which each subject's individual brain functional connection characteristics deviate from the normal range can be obtained. If the degree of deviation of the subject's functional connection characteristics exceeds the threshold, it indicates that an abnormality exists. Analyze the deviation pattern of the subject's brain functional connection characteristics to clarify which brain regions or networks have abnormal functional changes. Further, the subject's individual brain functional connection deviation score is input into the above-constructed classification model, and the classification model will return a result (ADHD or normal). Combined with the deviation pattern of the subject's individual brain functional connection characteristics and the results of the classification model, it is finally determined whether the subject is an ADHD patient, thereby realizing the identification of ADHD patients.

[0085] The following alternatives can achieve the same purpose of the invention without departing from the scope and spirit of the invention:

[0086] 1) Alternative data preprocessing methods: Image data preprocessing based on DPABI can be replaced by software based on SPM, FSL, fMRIPrep, etc. to also achieve preprocessing of functional magnetic resonance imaging data.

[0087] 2) Alternative brain functional connectivity matrix construction methods: Schaefer brain function atlas can be replaced by other high-resolution, high-credibility templates, such as Brainnetome, Glasser, Power templates, etc.; the calculation method of functional connectivity can be replaced by wavelet coherence, etc., to achieve the same construction of the functional connectivity matrix.

[0088] 3) Alternative modeling methods: Hierarchical Bayesian regression can be used as an alternative to normative modeling, which also uses Bayesian statistical methods to distinguish the sources of variation and is applicable to multi-center brain imaging data.

[0089] 4) Alternative classification and discrimination algorithms: SVM classification can be replaced by machine learning algorithms such as XGBoost, random forest, decision tree, etc. to achieve the same classification and discrimination of ADHD patients.

[0090] In summary, the present invention provides an attention deficit hyperactivity disorder identification system based on the developmental norms of brain functional connectomes. For the first time, a norm of brain functional connectome characteristics changing with age is constructed for children and adolescents. The Bayesian spline regression method used can model nonlinear distributions more accurately while shortening the calculation time. Therefore, it can be efficiently expanded to large brain imaging data queues and flexibly applied to new sites not included in the training samples. Based on the constructed developmental norms, the degree of deviation of ADHD patients from the normal range can be characterized at the individual level. The obtained individualized brain functional connection development deviation index contains brain function development information, which improves the classification effect of ADHD patients and controls compared with the original functional connection value. The present invention provides ADHD patients with quantitative, individual-level new brain imaging biological markers, thereby further revealing the neural heterogeneity mechanism of ADHD and promoting the accurate diagnosis and treatment of ADHD.

Claims

1. A system for identifying attention deficit hyperactivity disorder based on the developmental norms of brain functional connectome, characterized by: It constructs the developmental norms of the brain functional connection group and uses the obtained individual developmental deviation index of brain functional connections to classify and identify attention deficit hyperactivity disorder.

2. The attention deficit hyperactivity disorder identification system based on the brain functional connectome development norm according to claim 1, characterized in that: The brain functional connectome developmental norm is constructed using Bayesian spline regression.

3. The attention deficit hyperactivity disorder identification system based on the brain functional connectome development norm according to claim 1, characterized in that: It includes the following functional modules: a. Data acquisition module: used to collect demographic and magnetic resonance imaging data of subjects; b. Data preprocessing module: connected to the data acquisition module, used for correction, registration, noise reduction, smoothing and filtering preprocessing of the collected resting state functional magnetic resonance imaging data; c. Brain functional connection group calculation module: connected to the data preprocessing module, generating a functional connection matrix from the preprocessed resting state functional magnetic resonance imaging data, and calculating the functional connections of brain regions and brain networks; d. A brain functional connectome developmental norm construction module: connected to the brain functional connectome calculation module, in the normal development control group, using Bayesian spline regression to construct the norm of each brain region and brain network functional connection characteristics with age, based on the developmental norm, calculate the degree of deviation of each functional connection characteristic of each attention deficit hyperactivity disorder patient from the norm, and obtain the individual developmental deviation score; e. Attention Deficit Hyperactivity Disorder (ADHD) Identification Module: It is connected to the brain function connection group development norm construction module, and according to the obtained brain function connection development deviation score, a support vector machine algorithm is used to construct a classification model for ADHD patients and normal development controls, and the brain function connection pattern with individual extreme positive deviation and negative deviation is combined to determine whether the patient is an ADHD patient.

4. The attention deficit hyperactivity disorder identification system based on the brain functional connectome development norm according to claim 3, characterized in that: The data acquisition module is used to collect demographic information, T1 structural magnetic resonance imaging and resting-state functional magnetic resonance imaging data of multi-center attention deficit hyperactivity disorder patients and normal development control subjects; The data preprocessing module is for image data preprocessing. The functional magnetic resonance imaging data of all data sets are preprocessed using the brain imaging data processing and analysis toolbox based on the MATLAB platform; Brain functional connectome calculation: used to analyze and understand the functional coordinated activities or connection patterns between different brain regions, calculate the correlation of time series between different brain regions and networks, and obtain brain functional connectome characteristics; The developmental norms of brain functional connectomes were constructed in a control group with normal development. The data set was divided into 10 parts using a stratified 10-fold cross-validation method. The proportion of data from each center in each part was the same as that of the entire data set. One part was selected as the validation set each time, and the remaining 9 parts were used as the training set, and this cycle was repeated 10 times. Males and females were modeled separately to reduce the impact of gender differences on the model. Based on the training set data, Bayesian spline regression was used to characterize the changing pattern of functional connectivity characteristics of each brain region and brain network with age. Bayesian spline regression introduced cubic B-spline transformation into the general Bayesian linear regression model to capture nonlinear effects, and the Powell method was used to fit and optimize hyperparameters. Explained variance, EV, mean standardized log loss (MSLL) were used to evaluate the norms. The individual deviation score calculation method is as follows: based on the established brain functional connectivity norm, the degree to which each brain functional connectivity feature d of each subject n deviates from the norm is statistically estimated, that is, the individual deviation score Z nd , and its calculation formula is: Among them, y^ nd is the predicted functional connectivity value, y nd is the true functional connection value, using noise σ 2 d Variance (reflects variability in the data) and uncertainty variance (σ 2 * ) d Normalize.

5. The attention deficit hyperactivity disorder identification system based on the brain functional connectome development norm according to claim 4, characterized in that: The image data preprocessing method includes: removing the first 5 time points; time layer correction; head motion correction; image registration to T1 structural image and then to MNI standard space; regression Friston-24 head motion parameters, linear drift, white matter signal, cerebrospinal fluid signal and whole brain average signal; smoothing with a 6mm Gaussian smoothing kernel; 0.01-0.1Hz bandpass filtering.

6. The attention deficit hyperactivity disorder identification system based on the brain functional connectome development norm according to claim 4, characterized in that: The method for calculating the brain functional connectome comprises the following steps: S1: The cerebral cortex was divided into 400 regions of interest (ROIs) using the Schaefer brain functional atlas; S2: Calculate the Pearson correlation coefficient between the time series of these 400 ROIs and perform Fisher-z transformation to obtain a 400×400 functional connectivity matrix; S3: Calculate the average functional connectivity value of each ROI [the sum of the functional connectivity between the ROI and other ROIs / (total number of ROIs-1)] to obtain 400 brain region-level functional connectivity values; S4: According to the Yeo brain functional network atlas, these 400 ROIs were assigned to 7 brain networks, and the functional connectivity within and between networks was calculated, with a total of 28 [7+7×(7-1) / 2] functional connectivity values; the intra-network functional connectivity of each brain network was the average of the functional connectivity between the ROIs within the network, and the inter-network functional connectivity was the average of the functional connectivity between the ROIs of two brain networks.

7. The attention deficit hyperactivity disorder identification system based on the brain functional connectome development norm according to claim 4, characterized in that: The steps of constructing the norm are as follows: (1) The normal control data set was divided into 10 parts, and the proportion of data from each center in each part was the same as that of the entire data set. One of the parts was selected as the validation set each time, and the remaining 9 parts were selected as the training set. This cycle was repeated 10 times. Based on the training set data, a cubic B-spline basis function expansion was used for the independent variable age. φ(x) represents this expansion, where Φ is the N x K matrix containing the basis function expansion of all subjects. The single functional connectivity value y is assumed to be a linear combination of the B-spline basis function transformation plus a noise term: y=w T φ(x)+∈ s where w is the estimated weight vector, ε s is the Gaussian noise distribution of site s, with mean 0 and precision β s ; All noise accuracies from different sites will be combined in a vector β, allowing the noise accuracies to vary between sites to accommodate inter-site variability, as well as site-specific intercepts; The dummy variable site regressor was added to the final basis expansion matrix of the B-spline to make it a random intercept model; (2) Use Bayesian linear regression to give model parameters Set a Gaussian prior, where α is a weight-dependent hyperparameter; assume that the mean of the Gaussian prior is zero and the precision matrix is ​​Λ α , using the isotropic precision matrix Λ α = αI; the Gaussian prior choice allows the calculation of the posterior distribution of w in closed form: The posterior for each subject was obtained using standard posterior derivation: A=Φ T L β F+L α The Powell method is used to fit the optimal hyperparameters α and β; Powell is a derivative-free method that can fit hyperparameters faster. The resulting predicted distribution is: This predicted distribution reflects the changing pattern of brain functional connectivity characteristics with age, thereby obtaining the developmental norms of each brain region and network functional connectivity; the model was evaluated using the above-mentioned validation set data, and the explained variance (EV) and mean standardized log loss (MSLL) evaluation indicators were calculated to evaluate the fitting effect of the norm.

8. The attention deficit hyperactivity disorder identification system based on the brain functional connectome developmental norm according to claim 3, characterized in that: The method for identifying attention deficit hyperactivity disorder described in step e specifically comprises the following steps: S1: Using patients with attention deficit hyperactivity disorder as a test set, predicting the brain functional connectivity values ​​of patients with attention deficit hyperactivity disorder based on the above-constructed brain functional connectivity norm, calculating the individual deviation score of each brain functional connectivity feature of each patient with attention deficit hyperactivity disorder, defining the absolute value of z greater than 2 as an extreme deviation, in order to identify brain regions and networks showing extreme positive and negative deviations; S2: Combine the deviations of all brain regions and network functional connectivity to analyze the overall deviation pattern of patients with attention deficit hyperactivity disorder, and use extreme value statistics to analyze the most extreme deviation brain regions and networks of individuals; S3: The individual deviation scores of brain functional connectivity between patients with ADHD and normal development controls were used as input features to construct a support vector classification model to discriminate patients with ADHD.

Images